2021-05-31-CIDOB-Twitter在2021年对公众舆论和欧盟外交事务的影响（英文）_16页_760kb

Overview of the Study
This research analyzes the European Union's diplomatic communication and public perceptions of its foreign relations using Twitter data. By employing text mining, sentiment analysis, and regression modeling, the study examines over 927,000 tweets from 2016 to 2020 focusing on EU relationships with China, India, and Russia.

Methodology

• A dataset of English-language tweets was scraped using Python, filtered for engagement (likes and relevance), and preprocessed using R with packages for text analysis.
• Methods included sentiment (positivity/negativity scores) and emotion (anger, fear, joy, etc.) scoring via natural language processing (NLP), combined with linear regression to assess relationships and trends.
• Limitations noted: Twiter userbase is non-representative (young, educated, male), language bias (limited to English), and potential bot influence.

Key Findings

• Public Sentiment: Opinions are most positive toward the EU-India relationship, followed by EU-China, and least positive toward EU-Russia. Trends show increased attention during events like the Skripal poisoning and COVID-19 pandemic.
• Verbal Politeness Effect: EU official Twitter accounts use significantly more positive language than the general userbase, with stronger effects for China and India relations.
• Temporal and Temporal Dynamics: Tweet volume spiked during geopolitical events; sentiment evolved over time, with marked negative shifts in EU-China perceptions post-2020.
• Content Analysis: Top keywords reveal trade and cooperation for China, green energy and partnerships for India, and geopolitical issues like Ukraine and sanctions for Russia.

Conclusions
The study demonstrates that Twitter data can serve as a reliable proxy for gauging public opinion on foreign affairs, highlighting the EU's strategy of diplomatic politeness. However, finding validity and perceived effects worldlings and a of language.

// End of summary.